AWS Machine Learning Specialty MLS-C01 — Study Guide
Practice as you learn. Every concept maps to real exam questions with verified answers in the Amazon exam hub, or drill the full exam with timed practice on ExamRoll.io.
This guide covers each MLS-C01 domain in depth. Pick a domain to go deep, or work through them in order.
The domains
- Data Engineering & Feature Engineering — Challenge: Convert streaming CSV click events into Parquet in S3 with schema evolution, produce reliable feature vectors for a recommendation
- Exploratory Data Analysis & Visualization — Challenge: Identify customer segments that are most likely to churn in the next six months and establish monitoring to detect if feature
- Modeling — Supervised & Unsupervised (Classical ML) — Challenge: Identify feasible supervised/unsupervised use cases and a low-cost pipeline for training initial models with limited labels and
- Deep Learning & Computer Vision — Challenge: Build and deploy an accurate, low-latency queue-counting model that respects data security (private S3, KMS) and runs on constrained
- Natural Language Processing & Speech — Challenge: Identify the most-discussed topics from noisy, often short comments while handling slang, emojis, and thousands of proper nouns
- Time Series & Forecasting — Challenge: Produce calibrated 30-day probabilistic demand forecasts per SKU-store pair that account for holidays and promotions, handle
- Training, Distributed Training & Hyperparameter Optimization — Challenge: Forecast thousands of time series with long histories and imbalanced importance (stockouts cost more than overstocks), minimizing
- Deployment, Inference & Serving (ML Implementation & Operations) — Challenge: Provide low-latency predictions on a rolling 10-minute event window, redact PII before model training, ensure SageMaker can read
- Security, Privacy & Compliance — Challenge: Ensure end-to-end encryption and private networking for training and hosting, restrict image pulls to approved roles, and
- MLOps, Monitoring, Labeling & Model Governance — Challenge: Deliver low-latency anomaly detection at remote sites when connectivity is intermittent, while collecting telemetry for centralized
Ready to practice?
- Browse every question with verified answers → — free, with explanations.
- Start timed practice tests on ExamRoll.io → — the full question bank, in 20+ languages.
Pass the whole exam — not just this question
You found this answer. Get every verified question and explanation in one place, and save hours of prep. Free to start.
Pass your exam →